Arteriovenous fistulas as adjuncts to venous bypass grafts.

A canine model of occlusive hindlimb venous hypertension was used to determine which of two different, clinically applicable, adjunctive arteriovenous fistulas (AVFs), sequential or peripheral, would augment flow in autogenous cross-femoral venous bypass grafts (CFBs) with the least alteration of hindlimb hemodynamics. Unilateral venous hypertension was produced by iliofemoral venous ligation in three groups of five dogs: group I, venous ligation only (controls); group II, venous ligation followed by CFB with a sequential AVF; and group III, venous ligation with CFB and peripheral AVF. Bilateral hindlimb venous and arterial pressures and flows, and graft flows, were measured preoperatively and for 4 h postoperatively. Insertion of a CFB eliminated the venous hypertension in all 10 bypass dogs. Graft flow was augmented by addition of the sequential AVF (to 1167 +/- 309 mL/min from 92 +/- 12.3 mL/min: p less than .0001). However, this was accompanied, both in the limb ipsilateral to the AVF and in the contralateral limb, by the return of significant venous hypertension (p less than .001) and significant reduction (even reversal) of femoral vein flow caudad to the CFB (p less than .005). In contrast, adding the peripheral AVF augmented graft flow (to 200 +/- 62 mL/min from 65 +/- 43.7 mL/min; p less than .0001), but did not elevate venous pressure or impair venous flow in either hindlimb. The data from this short-term canine model suggest that a peripheral adjunctive AVF may effectively augment CFB graft flow without the potential for detrimental effects on venous hemodynamics characteristic of the sequential AVF.